This is my study. The data is on my computer.
Dr. Gordon and others made a claim of scientific misconduct at the University of Toronto last year, at which point I prepared a report on exactly what she's claiming. That report was submitted to the dean of public health sciences and was submitted to an international committee that reviewed the study and came out entirely on my side.
Let me tell you a couple facts. In the first round of screening, there were 270 palpable cancers on the mammography side, and 274 palpable cancers in the control group. If we had shunted women to the mammogram group who had a palpable cancer, that number would be different. There were 270 in the mammogram group, and 274 in the non-mammogram group.
Second, I removed all those women from the first round. I removed all the women with a palpable cancer from the analysis and reran it. The hazard ratio is 1.01.
Third, if what they are saying is true, then death from the cancer excessive in the mammogram group should have occurred in the first five years. This is a 30-year study. When I looked at the annual rates of mortality in the 30 years of follow-up, there was no difference in year one, year two, year three, year four and year five.
What Dr. Gordon is alleging is that I should see an excess of breast cancer deaths from the people who had prevalent breast cancer in the first round of screening, at which point we should see a high rate in the first group.
If I remove all the palpable cancers, which you can do, then I can get a hazard ratio of one. Second, the concept of a palpable cancer being excluded is ridiculous. Let's put it this way. In the studies that Dr. Gordon claims are the ones that provide evidence in favour of mammography was a Swedish two-county study. That was a study where the randomization was in 16 blocks of counties in Gothenburg and Östergötland, Sweden.
What did they do? The invited half the women to have a mammogram, and they didn't invite the others. They were just followed. Following the cancer rates—